Help Using DESeq without replicates

[choseqid]

Hi April,

What exactly did you do with dispTable(cds)[CondA]?

Thanks.

Hi All,

Figures the first time I post a question I end up answering it myself.

My conditions were not named what I thought they were. I solved this by using dispTable(cds)[CondA]. With the incorrect condition name this evaluated to NA, thus causing the error.

Sorry for the stupid question, maybe this will help a similarly misguided person!

[IsBeth]

My pipeline in DEseq using data without replicates and 2 conditions is not working either. I thought that something was wrong with the dispersion. But no dispersion method works, so...maybe there's a problem with the preprocessing? Anybody got the same problem with DESeq?

[jeky82]

I have the same error … but I did not understand how solve it, can you help me?

[indugun]

raju

Hello,

I'm working with a data set that we have yet to replicate. I'm trying to take a cursory look at the differential expression, but I keep getting an error I don't understand. I don't know if it matters but I'm generating my count tables with HTSeq. Here is the error:

> table = read.delim(file="~/neuron_rna_seq/hypodermis_table.txt", header=FALSE, stringsAsFactors=TRUE)
> cds = newCountDataSetFromHTSeqCount(table, directory="~/neuron_rna_seq")
> cds = estimateSizeFactors( cds )
> sizeFactors( cds )
hypodermis hypodermis_unfiltered
0.3333333 3.0000000
> cds = estimateDispersions(cds, method="blind", sharingMode="fit-only")
Error in parametricDispersionFit(means, disps) :
Parametric dispersion fit failed. Try a local fit and/or a pooled estimation. (See '?estimateDispersions')
In addition: Warning message:
glm.fit: algorithm did not converge
#Because of this I use the local fit
> cds = estimateDispersions(cds, method="blind", sharingMode="fit-only", fitType="local")
> results = nbinomTest(cds, "hypodermis", "hypodermis_unfiltered")
#Error in question
Error in if (dispTable(cds)[condA] == "blind" || dispTable(cds)[condB] == :
missing value where TRUE/FALSE needed


I apologize if this is a basic question, I'm not an experienced user.

Thanks!
April

Useful
Thank you

[Pseudoknot]

It was useful, thanks!

[aprilw]

Answer to my own question

Hi All,

Figures the first time I post a question I end up answering it myself.

My conditions were not named what I thought they were. I solved this by using dispTable(cds)[CondA]. With the incorrect condition name this evaluated to NA, thus causing the error.

Sorry for the stupid question, maybe this will help a similarly misguided person!

[aprilw]

Hello,

I'm working with a data set that we have yet to replicate. I'm trying to take a cursory look at the differential expression, but I keep getting an error I don't understand. I don't know if it matters but I'm generating my count tables with HTSeq. Here is the error:

> table = read.delim(file="~/neuron_rna_seq/hypodermis_table.txt", header=FALSE, stringsAsFactors=TRUE)
> cds = newCountDataSetFromHTSeqCount(table, directory="~/neuron_rna_seq")
> cds = estimateSizeFactors( cds )
> sizeFactors( cds )
hypodermis hypodermis_unfiltered
0.3333333 3.0000000
> cds = estimateDispersions(cds, method="blind", sharingMode="fit-only")
Error in parametricDispersionFit(means, disps) :
Parametric dispersion fit failed. Try a local fit and/or a pooled estimation. (See '?estimateDispersions')
In addition: Warning message:
glm.fit: algorithm did not converge
#Because of this I use the local fit
> cds = estimateDispersions(cds, method="blind", sharingMode="fit-only", fitType="local")
> results = nbinomTest(cds, "hypodermis", "hypodermis_unfiltered")
#Error in question
Error in if (dispTable(cds)[condA] == "blind" || dispTable(cds)[condB] == :
missing value where TRUE/FALSE needed


I apologize if this is a basic question, I'm not an experienced user.

Thanks!
April